The Problem with Arabic IDN's

[ChrisP]

I've registered a few arabic domains and then ran into a problem, I don't know the arabic language much - but from what I do know, I know that every character has 3 (maybe 4) ways of being written. It all depends on whether the character is at the beggining, middle (between any other two characters) or end.

So here comes the problem; take the word Locksmith in arabic:

صانع الأقفال

It's made up of two seperate "words" so there is a space in between. We'll call the one on the left "word 1" and the one on the right "word 2".

But when you try to get rid of the space you get:

صانعالأقفال

A totally different word all together! The character that made up the end of "word 1" joins up with the first character of "word 2" !!

So, the question is; can Arabic IDN work? I mean, if so, then surely we can only register one "word" words (lol confusing myself now).

Any ideas, what am I missing ?

[thefabfive]

You are correct. Many times putting two Arabic words together will change them. Sometimes it doesn't (when the last character of the first word only has one form).

For this reason many of us have stuck with single word Arabic domains. I do have some two word Arabic domains but they don't change when put together.

I advice would be to stick with single word Arabic domains.

[ChrisP]

I was really hoping I was wrong. This seems like a huge fundmental flaw in the arabic IDN structure. I can see one of two things happening:

> Single word arabic IDN's will go up as there is less real-estate.

or

> Arabic IDN won't take off because it's so limited.

Has anyone purposed a work around? Would an ASCII dash be an option people would consider ? e.g. :

صانع-الأقفالى.com

[Rubber Duck]

Originally Posted by ChrisP

I was really hoping I was wrong. This seems like a huge fundmental flaw in the arabic IDN structure. I can see one of two things happening:

> Single word arabic IDN's will go up as there is less real-estate.

or

> Arabic IDN won't take off because it's so limited.

Has anyone purposed a work around? Would an ASCII dash be an option people would consider ? e.g. :

صانع-الأقفالى.com

There are potential solutions for this. Something will be sorted out, but don't try to double guess ICANN on this one. An ASCII dash may not even exist on an Arabic Keyboard, so don't go that route.

I have a few unmodified ones of which the most recent are:

xn--mgbab2abb6excl1b.com بغدادفنادق Baghdad hotels
xn--mgbaah5aj7iegm1ai.com القاهرةفنادق Cairo Hotels
xn--mgbaah3aclo3kmej3ae1d.com الاسكندريةفنادق Alexandria hotels
xn--mgbayf0a0ej3a6a.com رياضفنادق Riyadh hotels

I guess those work OK?

[555]

I cant say i know it is on all arabic keyboards, but seems that it is

[thefabfive]

xn--mgbab2abb6excl1b.com بغدادفنادق Baghdad hotels
xn--mgbaah5aj7iegm1ai.com القاهرةفنادق Cairo Hotels
xn--mgbaah3aclo3kmej3ae1d.com الاسكندريةفنادق Alexandria hotels
xn--mgbayf0a0ej3a6a.com رياضفنادق Riyadh hotels

The first three look good. I can't tell about the last one. It looks like it may change but maybe only to line up with the next letter. It does seem to "stick" with the following word whereas the first three don't.

[Rubber Duck]

Yes, the representation on top the key may look the same but in Arabic Mode does it generate the same Unicode Point. Don't forget spaces and hyphens in Arabic start on the right and finish on the left, so it is hard to see how it could be the same Unicode Point. Similarly, the space is not allowed in domains, but it could be that a right to left space would be permited. I don't know the answers, but if you are intending to spend big bucks here, I would make sure you do.

[burnsinternet]

Originally Posted by ChrisP

I was really hoping I was wrong. This seems like a huge fundmental flaw in the arabic IDN structure. I can see one of two things happening:

> Single word arabic IDN's will go up as there is less real-estate.

or

> Arabic IDN won't take off because it's so limited.

I vote for "Single word arabic IDN's will go up as there is less real-estate!"

Seriously, Arabic is difficult language to 'mine' - just ask anyone here. I am pretty happy with mine. I don't have as many as some out there. Several good single word .coms and two decent .nets (الفيديو.net and زيت.net).

I wasn't able to compete head-to-head with others, so I was sneaking in French and Russian domains during the Arabic and Thai goldrushes. Right now, quality is important. Keep looking.

If you find something you absolutely love and it is registered, see if the owner will sell. Although, the owner may love it, too. Tokyo.net in Japanese sold for $10,000 a while back and Thai domains are not cheap, either.

[jose]

What do you mean with diferent words althogether?

صانع الأقفال

and

صانعالأقفال

are the same words without the space. They just "look" diferent. Don't believe me? Copy paste them on Word and do a search for each of the single words on the mengled one. Word will find each one easily.

Even for single words, any word, it can have a lot of formats, long, medium, small.
Look on wikipedia for Arabic Script.

Word does this automaticaly on some installations. Could yet find what triggers this mengling. For instance, my MSWord 2003 with XP SP2 from my desktop doesn't joins the letters on the words but the one on the laptop does (when I save the files). Couldn't find what you diferent. On both I don't have arabic language installed.

About the Rubber Duck juctions what worries me is that Google gives no results.